Modification frequency density with euclid distance classify for weather forecasting

Abstract

Method of weather forecasting fuzzy time series has some drawbacks, weather forecasting needs many weather components, big data size and forecaster experience, one of which is very difficult if it is worn on large data whereas weather Forecasting using time series forecast has large data characteristics, high dimensions, and continuous also require many large components and amounts of data, this affects speed and accuracy in weather forecasts, but the science develops in tandem with the development of the fuzzy Forecasting method, the fuzzy forecasting method that can be worn on large data will facilitate the weather forecasting, in this paper will be explained methods of forecasting for large data that is using the Euclid distance to classify data, then using the frequency density partitioning modification method applied to the database that has been grouped. The data used was KW Hipel Al McLeod 1994 using year variables and temperature, data from 1782 to 1988, data covers 206 global average temperature data annually. The final solution of the weather forecasting fuzzy time series problem is to analyze the value of AFER and MSE, the value of this paper is AFER 0.0021 and MSE 0.0025, the smaller the AFER and MSE values indicate that the method is good enough to forecast the weather.